The discovery of intracellular steroïd
hormone receptors in the 1960s presaged the
present era of investigation of DNA-binding,
gene regulatory proteins, of which the
steroïd/thyroïd hormone receptor
superfamily is an excellent example. It is
therefore not surprising that steroïds are
identified in many quarters of modern biology as
agents that regulate gene expression. Another
area of steroïd action is related to their
lipophilic character and their effect on cell
surface events. For example, the efficacy of
synthetic and natural relatives of
progestational steroïds as anesthetics has
been known for many years.
However, the relatively high concentrations
of steroïds needed to produce such effects,
as well as ignorance about a molecular mechanism
of action, relegated such apparent membrane
actions to the back burner while pharmacologists
pursued more precisely defined molecular
interactions.
During the past few years, the momentum has
shifted with the discovery of molecular
interactions of steroïds with membrane
events. At the same time there has been a
reinvestigation of a host of rapid and possibly
membrane-based actions of natural and synthetic
steroïds on a variety of tissues. The
effects o fsteroïds on excitable tissues
and neurosecretory processes are most prominent,
but there are also physiologically relevant
actions of progestins on the maturation of
spermatozoa and of oocytes. A picture is
emerging that shows that membrane and
intracellular actions of steroïds work,
sometimes in concert, to produce shortand
long-term modulation of cellular events,
particularly within the nervous system. In fact,
there has been sufficient interest in this topic
for it to be treated in depth recently at a ClBA
Foundation Symposium .
Steroïds and the GABAA
receptor
The GABAA/benzodiazepine receptor Cl channel
complex is a member of the same ligand-gated ion
channel superfamily as the nicotinic
acetylcholine receptor/ sodium channel complex,
and may exist in a number of forms determined by
its subunit composition. The receptor contains
several functional domains, including a Cl-
channel, a GABA recognition domain and a
benzodiazepine recognition domain. Certain
steroïds positively modulate GABA-induced
Cl- flux in a manner that resembles that of the
barbiturates, although the steroïd
modulatory site is believed to differ from that
of the barbiturates. In general, the
steroïds that most effectively modulate Cl-
flux are A-ring reduced, or pregnane
steroïd. There are two known natural
steroïd metabolites that are effective
modulators of Cl- flux: allopregnanolone
5alpha-pregnan-3alphaOH-20-one or 3alpha-OH-DHP,
which is derived from progesterone; and
5alpha-pregnan-3alpha,21diol-20-one (THDOC),
which is derived from desoxycorticosterone. Both
produce their effects on Cl- flux with IC50
values in the 10-50 nm range (these steroïd
concentrations are only effective when GABA is
also present). Both metabolites have been
identified in blood and brain tissue and show
increased concentrations within minutes after
stress. THDOC levels rise during stress from
<1 nm to reach concentrations -10 nm, which
are within the sensitive range of the receptor.
Moreover, enzymes in the brain generate these
metabolites from the parent steroïds, and
local concentrations of the metabolites may
reach high nanomolar concentrations because of
the contribution from this local production. The
progesterone metabolite 3alpha-OH-DHP is also
found in females and its levels in blood vary
during reproductive cycles. Furthermore,
anxiolytic effects of at least one of these
metabolites, THDOC, have been detected after
administration to rodents, suggesting that
stressinduced elevations or fluctuations during
the reproductive cycles might produce behavioral
effects. This, however, remains to be
established. In addition to their anxiolytic
actions, these steroïds also have
antiepileptic activity.
Facilitation of the inhibitoryeffects of
GABAA receptor activation by steroïds
appears to be a likely mechanism for the
anesthetic actions of many of these same
steroïds, although perturbations of the
membrane bilayer may also play a role. In order
to define more precisely the role of the protein
structure of the GABAA receptor complex in the
actions of steroïds and other modulators,
these receptors have been expressed in cells
transfected with DNA containing sequences for
the various subunit types. Expression of alpha
and beta subunits results in receptors that
respond to GABA as well as bicuculline and
barbiturates; however, benzodiazepine
sensitivity is reported to be missing in
alpha-beta chimerae and conferred by the
inclusion of DNA for ¶ subunits in the
expression system. steroïd sensitivity
apparent when beta alone, or alpha plus beta or
alpha plus beta plus ¶ subunit DNA is expressed
Thus the receptor protein itself is a key factor
in determining steroïd sensitivity the
GABAA receptor. There some preliminary evidence
that different types of alpha subunit confer
differing degrees of steroïd sensitivity on
the resultant GABA receptor complex when
co-expressed with beta and ¶ subunit DNA.
Progestational steroïds and Ca
2+
Oocytes and spermatozoa exist in fluids that
contain reproductive hormones, and certain
aspects their maturation are regulated the
progestational steroïds. Studies on the
frog oocyte have revealed a mechanism of action
of progesterone at the oocyte surface that
causes increased levels of Ca2+ within the egg
cytoplasin. These effects of progesterone are
rapid, with onset latencies of 40-60s; they last
for 5-6 min. This Ca2+ mobilization is involved
the meiotic maturation of th oocyte.
The acrosome reaction maturational event in
spermatozoa that involves a proterone stimulated
influx of cellular Ca2+ similar to the events
occurring in the oocyte. Progesterone itself is
very effective in producing these effects, and
structure-activity profile of other
steroïds appears to differ from that of the
GABAA receptor activation described above,
although many key steroïds have not be
tested on both systems. The effects of
progesterone occur within 30-60 s and are due
influx of Ca+2+ from the extracellular space,
since they are blocked both by the chelator EGTA
and by the Ca 2+ channel antagonist lanthanum
ions. The effect of progestins on Ca2+ levels
inside cells is relevant to the actions of these
steroïds on neurotransmitter and
neurohormone release and such effects are
summarized in Table 1. Transmitter release
depends on Ca2+ and the modulation of Ca 2+
availability within the cell may represent a
general mechanisrn of action of such
steroïds at the cell surface.
Other rapid effects of
steroïds
The specific interactions of steroïds
with putative membrane receptor sites related to
the GABAA receptor and to Ca` flux have prompted
a re-examination of some old phenomena as well
as new investigations of nongenomic effects of
steroïds. As shown in Table I, there is an
impressive array of effects, some of which may
have physiological relevance, while others may
not. The higher the steroïd concentrations
that are required for effect, the less likely
they are to be achieved in vivo. For example,
the interference by 2-hydroxyestradiol of
dopamine and alpha1-adrenoceptor binding
involves extremely high steroïd
concentrations that are unlikely to be achieved
in vivo. Another consideration is whether the
effect itself occurs in vivo. The actions of
3alpha OH-DHP and THDOC on the GABAA receptor,
as well as the actions of progesterone on oocyte
and sperm maturation, are likely candidates.
Steroïds have a number of rapid effects
on neurosecretion . These actions occur at
concentrations that are physiologically quite
reasonable, and they are rapid in onset,
occurring as rapidly as measurements of release
can be made after adding them to the fluid
bathing the cells. A case can be made for the
actions of progestins on luteinizing hormone
releasing hormone (LHRH) and dopamine release
occurring in vivo.
Table 1 also notes extremely rapid effects of
steroïds on electrical activity of nerve
cells, occurring when applied locally. Methods
of application include iontophoresis and
pressure ejection, where experiments are carried
out in vivo, and application to the bath for in
vitro preparations. The onset latencies of these
effects are all within seconds of application,
which excludes any type of genomic action.
Structure-activity studies in at least one case
indicate a considerable degree of steroïd
specificity, which argues for a receptor or
recognition site at or close to the cell
surface.
In pursuit of membrane receptors
One of the most elusive goals has been to
obtain evidence with radiolabeled steroïds
for the existence of putative membrane receptor
sites. A number of studies showing membrane
binding sites for steroïds are summarized
in Table 1. A study by Towle and Sze employing a
centrifugation assay, produced suggestive
evidence for the existence of brain membrane
binding sites for glucocorticoids, androgens,
progestins and estrogens. A more recent study,
using a filtration assay, has demonstrated high
affinity membrane-associated receptors for
estrogens in the rat pituitary gland.
One of the reasons for the elusiveness of
membrane sites for steroïds may be their
low abundance. Availability of higher specific
radioactivity 125 I-labeled progestin has now
allowed the successful identification of binding
sites in brain membranes which have some of the
specificity expected of the sites that regulate
neurosecretion. The adsorption of steroïds
by membranes is another problem; this means that
nonspecific binding is inevitably very high and
makes extensive studies of membrane binding
extremely difficult.
In two instances, steroïds have been
immobilized on macromolecules in order to
demonstrate cell surface actions and receptors.
In one series of experiments, progestins
conjugated to serum albumin were used both for
the 125I labeled progestin binding studies to
brain membranes and to demonstrate the effects
of steroïds on neurosecretory processes. In
another study, estradiol conjugated to nylon
fibers was used as an affinity column to isolate
cells with membrane estrogen receptors. The
conjugation of steroïds to macromolecular
supports thus shows great promise in identifying
membrane steroïd receptors and in isolating
receptor-containing cells, as well as providing
more definitive evidence for membrane
steroïd actions mediating cellular
processes.
Genomic actions of steroïds
In excitable tissues such as brain, long-term
signaling by circulating hormones via genomic
mechanisms plays an important role in shaping
cell structure and function.
In the nervous system, the effects of
steroïds are usually confined to particular
groups of cells that contain
intracellular
steroïd receptors. The effects produced
through these receptors range from the induction
of key enzymes of neurotransmitter metabolism
and neurotransmitter receptors to induction of
synaptic and dendritic structure. Most of these
changes are reversible ones, and some have been
shown to vary during natural endocrine cycles
such as the estrous cycle of the female rat.
The effects of estrogens are among the best
studied. In basal forebrain cholinergic cells,
for example, estradiol induces the enzyme
choline acetyltransferase which is the key
enzyme for acetyl choline biosynthesis. In
ventromedial hypothalamus, estradiol induces
synthesis of receptors for the neuropeptide
oxytocin and also induces oxytocin formation in
the cells that appear to innervate these
receptors; as will be discussed below, these
events may have importance for estrogen
induction of female sexual behavior. Estradiol
also induces synthesis of progesterone receptors
in a number of brain regions, as w as in the
reproductive tract and pituitary, and this
induction is important for induction of sexual
behavior. Longterm estrogen induced increases in
5-HT1 receptors and GABAA receptors also occur
in certain brain regions, and increases in mRNA
for enkephalin in the ventromedial nuclei of
hypothalamus have been observed. All of these
effects take many hours or days to appear, and
occur only in brain areas with intracellular
estrogen receptors. Thus they appear to be
genomic effects. Perhaps the most surprising
finding has been that estradiol induces spines
on dendrites neurons within the hypothalamus and
hippocampus of the female rat; since spines are
occupied by synapses, this finding strongly
suggests that the hormone is regulating
synaptogenesis; electron microscope data support
this suggestion . Even more prising is that
spine densities these same dendrites rise and
throughout the four-day estrous cycle of the
female rat.
Other steroïds also produce long-lasting
and apparently genomic effects on neural tissue.
In addition to regulating the neuropeptide
systems that govern ACTH release from the
pituitary gland, glucocorticoids regulate a
number of enzymes and structural proteins
throughout the brain as well as having long-term
effects on neuronal survival and
destruction.
Another type of long-lasting steroïd
effect that alters tissue excitability is
illustrated by a recent study of the action of
androgen in muscle. Androgens act on myotubes
from frog muscles, which contain intracellular
androgen receptors, to exert a long-term
increase of acetylcholine-activated single
channel conductances, suggesting that the
steroïd induces a substance that modulates
channel behavior in the acetylcholine receptor.
This study emphasizes the importance of careful
analysis of how actions of a steroïd
produce a membrane effect, to distinguish
between genomic and nongenomicic mechanisms.
The most obvious way of distinguishing these
mechanisms is on the basis of time-course, with
rapid onset of effects being nongenomic and
possibly membrane mediated, and effects that are
slower in onset being genomic. The problern in
making a distinction lies in the time range of
minutes, where rapid genomic or long-lasting
membrane effects might be involved. To resolve
such cases, inhibitors of protein or RNA
synthesis may be useful, along with other means
of isolating the membrane from the rest of the
cell, such as patch clamping.
Genomic and non-genomic effects in
concert
Figure 3 summarizes some of the actions of
progesterone and its metabolites, at the
membrane level and via the genorne, emphasizing
the fact that metabolism of the steroïd by
enzymes in neurons, or on or in glial cells, may
be a regulatory mechanism for generating the
steroïd derivatives that have effects on
different membrane receptors. The gene products
that are regulated by progesterone acting via
intracellular receptors are not identified as
yet. However, their apparent importance for the
modulatory effects of progesterone on
reproductive behavior has been recognized. The
scheme shown in Fig. 3 is also applicable to
other steroïds such as those from the
adrenal cortex, where metabolism generates some
membrane-active steroïds and where genomic
actions of the parent steroïd are also
known.
Interaction and cooperation could
theoretically occur between genomic and
non-genomic steroïd responses.
Indeed,effects such as those of the progestins
on responses of neurons to glutamate and GABA
are independent of prior estrogen priming.
However, some membrane effects of
iontophoretically applied estradiol on neural
activity are influenced by the stage of the
estrous cycle. This implies a dependence on the
actions of ovarian hormones, possibly via
genomic mechanisms.
We have recently found that estradiol induces
oxytocin receptors in the ventromedial
hypothalamus of the female rat and that at least
some of these induced receptors, located in the
caudal ventromedial nuclei, will then respond
rapidly to the administration of progesterone.
Within less than an hour, progesterone causes an
apparent spread of the oxytocin receptor field
in a direction lateral to the cell bodies where
the receptors are induced. It appears that this
spread may be occurring along the dendrites of
the ventromedial nucleus neurons which project
laterally. Much to our surprise, progesterone
was able to produce the same effect in vitro in
previously freeze-dried sections prepared for
binding of the oxytocin receptor ligand. The
fact that this action occurs in vitro at 100 nm
progesterone and appears to be the same as the
effect of physiological doses in vivo argues
strongly that this is a physiologically relevant
membrane effect of the steroïd. It may
represent a rapid movement of receptors along
the dendrites or it may be an activation of
low-affinity receptors to a high-affinity form
capable of binding the ligand. As far as we
know, the steroïd specificity favors
progesterone over estradiol cholesterol and some
of the progestin metabolites that affect the
GABAA receptor, but further studies are
required. Whichever mechanism is involved, the
result of this action of progesterone appears
causally linked to the induction of sexual
behavior in the female rat, based upon
behavioral and anatomical studies.
Steroïd hormones have long been regarded
as acting on cells via totally different sites
and mechanisms of action from those of
neurotransmitters. Neurotransmitters, as well as
hormones that act on the surface of cells, were
traditionally thought to influence only
short-term responses at the cell membrane. The
discovery of increasing numbers of second
messengers and processes linked to them has led
to the realization that they affect many
intracellular processes, including changes in
gene expression. Now, with the recognition of
the actions of steroïds on membranes as
described in this article, the distinction has
broken down further. That is, the actions of
steroïds appear to involve the cell
membrane as well as the genome, and (although it
has not been demonstrated) it is even
conceivable that actions of steroïds at the
cell surface might be able to trigger changes in
gene expression indirectly.
We need to know much more about membrane
effects and the steroïds that cause them
before we can fully appreciate the connections
between their actions at the membrane and on the
genome. First, it is important to know whether
there are unifying processes that underlie at
least some of the membrane effects summarized in
Table 1. For example, how many of these actions
may be explained by steroïd-induced
alterations in intracellular Ca2+ ? Furthermore,
is the action of steroïd on the GABAA
receptor an isolated example, or do other
complex receptors such as those for
acetylcholine or NMDA also contain steroïd
recognition sites? Secondly, it is essential to
reexamine the extensive work on steroïd
metabolism in brain to see how many natural
metabolites of steroïds have membrane
activity, as the A-ring reduced metabolites of
progesterone and adrenal steroïds have on
the GABAA receptor. Moreover, the known membrane
effects of steroïds must be compared across
the same group of steroïds in corder to
understand fully similarities and differences in
structure-activity profiles. Thirdly, it is
important to find out whether all membrane
associated receptor sites can be ascribed to
protein structures in the membrane, or whether
there is a separate class of 'recognition' sites
defined largely by the cornposition of menbrane
lipids. Finally, development of new psychoactive
drugs may benefit from considering the
membrane-active steroïds either themselves
or as starting points for the creation of new
synthetic compounds. The diversity of
structure-activity profiles already evident in
the case of the GABAA receptor makes this area a
fertile one for future research.
